python - Construct DatetimeIndex if you only have year -

i have data of structure

     country  year        pop 606  algeria  1966  12339.140 730  algeria  1968  13146.267 793  algeria  1969  13528.304 856  algeria  1970  13931.846 924  algeria  1971  14335.388 

now want create first-differences per country based on year (difference per year). if weren't interval concern, i'd along lines of

df.sort(['country', 'year']).set_index(['country', 'year']).diff() 

instead, guess have convert year to_datetime() first. there simple way create datetime column contains years only? , there different more natural approach create differences on time?

you do

df.set_index(df.year.map(lambda x: datetime.datetime(x, 1, 1))) 

that uses concept of left-open intervals.

another possibility is

df.set_index(df.year.map(pd.period)) 

both return equally well-defined indexes, in latter case might output of df.diff() better since states year.